A comparative analysis of vintage and non-vintage capital growth models

Almost all economic growth models assume that technology is disembodied, not inherent in the capital stock. Under this assumption, capital is homogeneous and access to technology is not linked to capital investment. This assumption contradicts simple intuition and empirical evidence showing that the majority of technology is embodied in the capital stock. Embodied technology leads to a heterogeneous capital stock in which a unit of capital is defined by when it is produced, its vintage. This dissertation examines the possible consequences of making the traditional simplifying assumption of disembodied technology, by comparing the results of vintage and non-vintage capital models.; The first two papers of the dissertation examine neoclassical growth models. These are models in which the growth rate of technology is exogenous. The three standard vintage capital models are examined, putty-putty, clay-clay, and putty-clay. The first paper focuses on convergence issues and shows that there are considerable differences in the transitional dynamics of vintage versus non-vintage models.; The second paper focuses on the mathematical methods used to obtain the solutions of paper one. It is shown that the putty-putty and clay-clay models can be written as convex programming problems, and that therefore interior point methods will be well defined and converge. It also includes the formulation of the dual problems to the putty-putty and clay-clay models. The steps involved in the formulation lead to economic results utilizing the shadow prices.; The growth of technology is often the result of dedicated research and development. Therefore, the third paper of the dissertation focuses on two-sector models. In these models, technology will only increase if the central planner allocates scarce resources to the research sector. Two different models are examined. The first model utilizes “AK” production in the research sector, while the second utilizes Cobb-Douglas production. The AK model is the first model to demonstrate that vintage and non-vintage capital versions of models can have different long-run growth rates. In both models, embodied technology leads to a higher savings rate and less labor allocated to research.